Back to Search
Start Over
7. Comparison of two deformable registration software using an in-house 'deformable phantom'
- Source :
- Physica Medica. 32:344
- Publication Year :
- 2016
- Publisher :
- Elsevier BV, 2016.
-
Abstract
- Introduction The multimodality image fusion, widely used in radiotherapy to precisely contour the target volumes, opened the field to deformable registration, as significant anatomical and positional modifications may occur between acquisitions. In this work, a heat deformable 3D phantom, with two rigid well defined lesions, was conceived to quantify the performances of Velocity and Smart Adapt in their ability to segment structures on deformed images. Material and methods The phantom’s deformable body (BODY) is made from wax (deformable at 40 °C) and includes two rigid elements of Aquaplast (deformable at 60°C) in its extremities (Fig. 1). The volumes of the structures were 10.78cc, 18.9cc and 466.8cc respectively for PTV-A, PTV-B and BODY. The initial distance between the centers of PTV-A and PTV-B was 10.6 cm. At first, it was scanned on its initial form (CTinitial), then undergone two deformations, having one CT imaging for each deformation (CT-1deform, CT-2deform). The PTV-B was shifted from PTV-A 1.2cm after the first deformation and 4.7cm after the second deformation. Reference contours BODY, PTV-A and PTV-B were manually contoured on each CT, with 3 physicist drawing PTVs to estimate intra-observer variability. The CTinitial was deformed to CT-1deform and CT-2deform using both Smart Adapt and Velocity. Different metrics were used to compare manual and segmented structures. Results In the table below are presented the metrics results as a mean value for PTV-A & PTV-B. Download : Download high-res image (112KB) Download : Download full-size image
- Subjects :
- Computer science
business.industry
Mean value
Biophysics
Planning target volume
General Physics and Astronomy
General Medicine
Deformation (meteorology)
Multimodality image fusion
Imaging phantom
Software
Computer graphics (images)
Radiology, Nuclear Medicine and imaging
Computer vision
Artificial intelligence
Ct imaging
business
Subjects
Details
- ISSN :
- 11201797
- Volume :
- 32
- Database :
- OpenAIRE
- Journal :
- Physica Medica
- Accession number :
- edsair.doi...........d4e6aa1e4b7b164fe6291246e111f71a
- Full Text :
- https://doi.org/10.1016/j.ejmp.2016.11.058